Effect of Hydrophilic Swellable Polymers on Dissolution Enhancementof Carbamazepine Solid Dispersions Studied Using ResponseSurface MethodologySubmitted: June 8, 2006; Accepted: September 25, 2006; Published: April 6, 2007

Yogesh Rane,1 Rajshree Mashru,1 Mayur Sankalia,1 and Jolly Sankalia1

1Pharmacy Department, Faculty of Technology and Engineering, The M. S. University of Baroda, Kalabhavan, Vadodara -390 001, Gujarat, India

ABSTRACT

The objective of this work was to study dissolution en-hancement efficiency and solid dispersion formation abilityof hydrophilic swellable polymers such as sodium carbox-ymethyl cellulose (Na-CMC), sodium starch glycolate (SSG),pregelatinized starch (PGS), and hydroxypropylmethyl cel-lulose (HPMC) with carbamazepine using 32 full factorialdesign for each of the polymers. Solid dispersions of carba-mazepine were prepared using solvent evaporation methodwith around 70% solvent recovery. The independent var-iables were the amount of polymer and organic solvent. Thedependent variables assessed were percentage drug dis-solved at various time points and dispersion efficiency (ie,in terms of particle size of solid dispersion). Solid disper-sions were evaluated for percentage drug dissolved, wett-ability, differential scanning calorimetry, scanning electronmicroscopy, and angle of repose. Multiple linear regressionof results obtained led to equations, which generated con-tour plots to relate the dependent variables. Similarity factorand mean dissolution time were used to compare dissolu-tion patterns obtained in distilled water and simulated gastricfluid United States Pharmacopeia (USP) XXVI of pH 1.2.Maximum drug dissolution was obtained with polymer or-der Na-CMC9SSG9PGS9HPMC. Particle size of drug wasreduced ~10–15, 3–5, 5–7, and 10–25 times in Na-CMC,SSG, PGS, and HPMC solid dispersions, respectively;whereas wettability of solid dispersions was found in theorder of Na-CMC9HPMC9PGS9SSG. Angle of repose wasfound to be in the range of 29° to 35° for all solid disper-sions, which shows good flowability characteristics. HPMCshowed increase in drug dissolution up to an optimized lev-el; however, further increase in its concentration decreaseddrug dissolution.

INTRODUCTION

The rate of oral absorption of poorly soluble drugs is oftencontrolled by their dissolution rate in the gastrointestinaltract.1 Thus solubility and dissolution rate are the key deter-minants of oral bioavailability, which is the concluding pointdrawn for fate of oral bioavailability.2,3 Carbamazepine (CBZ)is a widely prescribed antiepileptic drug having poor watersolubility (~170 mg/L at 25-C).4 Because of having poorwater solubility, its absorption is dissolution rate limited,which often results in irregular and delayed absorption.5

For improvement of solubility and dissolution rate of poorlysoluble drugs, numerous commercially viable techniquessuch as liquisolid, in which drug in solution state or dis-solved drug is adsorbed over insoluble carriers,6-8 nanomorph,a patented technology by Soliqs for controlled crystalliza-tion of drug,9 in situ micronization,10,11 and coprecipitationusing antisolvent,12 are available. Surfactants can also beused in formulations to improve wettability and solubilityof many lipophilic substances.13 Micronization of drug isnot preferred because micronized product has the ten-dency of agglomeration, which leads to decreased effectivesurface area for dissolution.14 But ahead of all, solid dis-persion is the most promising method to formulators be-cause of its ease of preparation, ease of optimization, andreproducibility.15-18 Poorly soluble drugs are dispersed in aninert hydrophilic polymer or matrix by melting, solutionformation, or solvent melting to yield solid dispersion.15,17

Usually, solid dispersions (SDs) are prepared with watersoluble low melting point synthetic polymers such as poly-vinylpyrrolidone (PVP), mannitol, or polyethylene glycols(PEGs).19 These polymers show superior results in drugdissolution enhancement, but the amount of these polymersrequired is relatively large, around 1:2 to 1:8 (drug/polymer)ratio.20 In certain similar experiments it has been observedthat, PVP and PEG get dissolved first in dissolution me-dia (owing to their high water solubility) leaving the drugback in undissolved state. In such case, though the drug isin controlled crystallization state or amorphous state, the

KEYWORDS: Solid dispersion, factorial design, similarityfactor, mean dissolution time, carbamazepine, multiplelinear regression, simulated gastric fluid.

AAPS PharmSciTech 2007; 8 (2) Article 27 (http://www.aapspharmscitech.org).

E1

Corresponding Author: Rajshree Mashru, PharmacyDepartment, Faculty of Technology and Engineering, TheM. S. University of Baroda, PO Box 51, Kalabhavan,Vadodara - 390 001, Gujarat, India. Tel: +91-265-2434187/2434188; Fax: +91-265-2423898/2418927; E-mail:rajshreemashru@yahoo.com

polymers are unable to provide wetting ability to the drugparticles. In such cases, there may be the possibility of rapidreversion of amorphous drug to the more stable crystallinestate in presence of small amount of plasticizers such aswater.21

Literature survey reveals that certain hydrophilic swellablepolymers such as sodium carboxymethyl cellulose (Na-CMC), sodium starch glycolate (SSG), and pregelatinizedstarch (PGS) have still been unexplored for their potentialto form solid dispersion in order to improve dissolution prop-erties of poorly soluble drugs. For this reason, in the presentwork, water-swellable polymers or normal excipients of soliddosage forms were used. These polymers were supposed tohold the drug in intimate contact with water (owing to theirwater retention potential) and increase its wettability. Soliddispersions were prepared with modified solvent evaporationtechnique.

Full factorial experimental design is one of the best tools tostudy the effect of different variables on the quality deter-minant parameters of any formulation. In the present study,independent variables were assigned to the amount of poly-mer and the amount of solvent at 3 different levels, whereasresponses or dependent variables for them were assigned topercentage drug dissolved at various time points and disper-sion efficiency (ie, in terms of particle size of solid disper-sion). Q30 (ie, percentage drug dissolved at 30 minutes) wasdetermined for 2 different dissolution media (ie, distilledwater and SGF without enzymes [USP XXVI]). The multiplelinear regression (MLR) analysis of results led to equationsthat adequately described the influence of the independentvariables on the selected responses. Polynomial regressionequations and contour plots were used to relate the depen-dent variables. As part of the optimization process, themain effects, interaction effects, and quadratic effects of theamount of polymer and amount of solvent on percentagedrug dissolved of solid dispersion were investigated.

Gohel and Panchal22,23 recently proposed a similarity fac-tor (Sd) for the comparison of dissolution profiles, whichis more simple and flexible than similarity factor (f2) be-cause data can be expressed either as the amount of drugdissolved or as the percentage of drug dissolved. Anotheradvantage is that, unlike the similarity factor (f2), linearinterpolation can be used to accurately express the results.The dissolution profiles of optimized batch in distilled wa-ter and simulated gastric fluid (SGF) without enzymes ofpH 1.2 were compared using similarity factor (Sd).

MATERIALS AND METHODS

Chemicals

Analytical grade chemicals were used as received. Carbama-zepine was received as gift from Relax Pharmaceuticals Ltd,

Vadodara, India. Polymers were purchased from S. D. FineChemicals Ltd, Mumbai, India. Deionized double-distilledwater was used throughout the study.

Preparation of Solid Dispersion

Solid dispersions of all the polymers were prepared by mod-ified solvent evaporation method,10 wherein drug was dis-solved in acetone at its saturation solubility with continuedstirring up to 30 minutes. Polymer was suspended in suffi-cient amount of water (up to wet mass of polymer). Thedrug solution was poured at once into polymer suspension.The entire solvent was evaporated under reduced pressureat 60-C to 70-C with Rotavapor (Heidolph, Germany) withsolvent recovery. The recovered solvent was used for thenext batch. The solid dispersion was obtained in a flask,which was dried at 70-C to 80-C and stored in a desiccatorfor 24 hours.

Solubility Determination

For the determination of solubility of carbamazepine, ex-cess material was placed in contact with 7 mL of solvent insealed glass tubes. The tubes were shaken on a vortex mixerand were maintained at 25-C for 24 hours. The saturatedsolution was centrifuged and the supernatant was filteredthrough 0.45-µmWhatman filter paper (Whatman Ltd, Mid-dlesex, UK) diluted suitably with water and analyzed by UVspectrophotometer at 285 nm (model UV-1601, UV-Visiblespectrophotometer, Shimadzu, Kyoto, Japan).

Experimental Design

To study all the possible combinations, 3 levels full factorialdesign (32) was constructed and conducted in a fully ran-domized order.24 The dependent variables measured werepercent drug dissolved at various time points and particlesize of solid dispersion at three different levels. Independ-ent variables of the 32 full factorial design with their codedand actual values are shown in Table 1. The range of a factorwas chosen in order to adequately measure its effect on theresponse variables. This design was selected as it providessufficient degrees of freedom to resolve the main effects aswell as the factor interactions. MLR analysis was used tofind out the control factors that affects significantly on re-sponse variables.

Differential Scanning Calorimetry

Differential Scanning Calorimetry (DSC) thermograms ofpure CBZ, CBZ: Na-CMC, CBZ: SSG, CBZ: PGS, CBZ:HPMC solid dispersions were measured using differen-tial scanning calorimeter (DSC 60, Shimadzu) previously

AAPS PharmSciTech 2007; 8 (2) Article 27 (http://www.aapspharmscitech.org).

E2

calibrated using indium. The samples ~2 to 3 mgs wereaccurately weighed into solid aluminum pans with sealsand crimped. Reference pan was an empty sealed aluminumpan. The measurements were obtained at a heating rate of10-C/min with purging of dry nitrogen at a constant rate of20 mL/min.

Scanning Electron Microscopy

To observe the surface morphology of CBZ and its SDswith Na-CMC, SSG, PGS, and HPMC scanning electronmicroscopy (SEM) studies were performed using Jeol JSM-5610 LV (Jeol Corp, Tokyo, Japan). CBZ and SDs wereindividually glued on the brass sample holder with the help

Table 1. Matrix of Independent Variables and Responses for 32 Factorial Design for Each Polymer*

Variables With Levels† Response Values

Batch X1 X2

Q30 for %Drug Dissolved

in Water

Q30 for %Drug Dissolved

in SGF

Particle Sizeof OptimizedBatch (μm)

Angle of Reposefor OptimizedBatch (o)

Wettability Timefor OptimizedBatch (minutes)

Polymer: Na-CMCA1 –1 –1 66.58 64.25 15–17 29.43 03.15A2 –1 0 70.21 68.45A3 –1 1 71.35 67.21A4 0 –1 82.64 76.85A5 0 0 87.63 83.54A6 0 1 88.54 85.65A7 1 –1 93.58 91.14A8 1 0 95.21 93.47A9 1 1 95.64 94.69

Polymer: SSGB1 –1 –1 64.25 61.26 138–158 34.56 04.49B2 –1 0 67.51 65.24B3 –1 1 67.85 65.36B4 0 –1 76.36 74.21B5 0 0 78.21 75.29B6 0 1 78.62 76.91B7 1 –1 83.64 80.24B8 1 0 84.25 82.91B9 1 1 85.65 83.16

Polymer: PGSC1 –1 –1 65.36 63.59 10–200 30.65 03.86C2 –1 0 67.63 65.37C3 –1 1 68.62 66.69C4 0 –1 75.54 73.16C5 0 0 75.69 73.26C6 0 1 77.25 74.36C7 1 –1 76.58 75.91C8 1 0 82.01 78.12C9 1 1 82.35 79.19

Polymer: HPMCD1 –1 –1 65.68 62.46 2–79 33.25 03.56D2 –1 0 66.54 64.23D3 –1 1 66.25 65.28D4 0 –1 68.35 66.47D5 0 0 73.52 70.25D6 0 1 74.56 72.19D7 1 –1 70.21 71.25D8 1 0 69.25 69.69D9 1 1 69.58 70.58

*Na-CMC indicates sodium carboxymethyl cellulose; SSG, sodium starch glycolate; PGS, pregelatinized starch; and HPMC, hydroxypropylmethylcellulose.†Independent variables levels: low (–1), medium (0), high (1); amount of polymer (1 g, 2 g, 3 g); and amount of solvent (150 mL, 200 mL, 250 mL).

AAPS PharmSciTech 2007; 8 (2) Article 27 (http://www.aapspharmscitech.org).

E3

of double-sided adhesive tape. The images were captured atan excitation voltage of 15 kV at varying magnificationsfrom original magnification ×50 to ×650.

In Vitro Dissolution Study

The dissolution study was performed using 2 media, dis-tilled water and SGF without enzymes. Accurately weighedamount of solid dispersion, containing equivalent 100 mgof pure drug was placed in basket of USP XXIV dissolu-tion apparatus (Type I, TDT-06P, Electrolab, Mumbai, India)with 900 mL deaerated dissolution medium. Deaeration ofdissolution media was done by ultrasonication (Ultrasonics -2.2, Mumbai, India) of dissolution medium for 15 minutes.The dissolution apparatus was run at 100 rpm at constanttemperature 37-C ± 1-C. Samples (5 mL) were withdrawnat 0, 5, 10, 15, 20, 25, and 30 minutes, filtered through0.45-µm Whatman filter paper, diluted suitably and anal-yzed spectrophotometrically at 285 nm (model UV-1601UV-visible spectrophotometer, Shimadzu). An equal volumeof fresh dissolution medium kept at the same temperaturewas added to maintain the sink conditions. The absorbancevalues were transformed to concentration by reference toa standard calibration curve obtained experimentally (r2 =0.9998). The dissolution test was performed in triplicate foreach batch.

Similarity Factor Sd

The similarity factor Sd is defined by Equation 1.

where n is the number of data points collected during the invitro dissolution test; AUCwt and AUCGt are the areas undercurves of the dissolution profiles of the solid dispersion indistilled water and SGF without enzymes, respectively, attime t. For the dissolution profiles in distilled and SGF with-out enzymes, to be identical, the Sd value should be zero.

22,23

Wettability Studies

Pure drug, weighing ~1 g was placed in sintered glass fun-nel of 27 mm internal diameter. Bridge was formed at the

neck of the funnel with the help of cotton plug. The funnelwas held in upright position in a beaker filled with watersuch that the water level in the beaker just touched the cottonplug. Methylene blue powder was layered over the surface ofpure drug in the funnel. The time required to raise the waterthrough the drug till wetting of methylene blue powder oc-curred was recorded.25 The procedure was followed for allthe SDs.

Angle of Repose

To get an idea about flowability properties of the solid dis-persions, angle of repose for all the batches of experimentaldesign was determined. If the angle exceeds 50-, the mate-rial will not flow satisfactorily, whereas materials havingvalues near the minimum flow easily and well. The rougherand more irregular the surface of the particles, the higher isthe angle of repose.26 The angle of repose was measured bypassing SD through a sintered glass funnel of internal di-ameter 27 mm. on the horizontal surface. The height (h) ofthe heap formed was measured with a cathetometer, and theradius (r) of the cone base was also determined. The angleof repose (Φ) was calculated from Equation 2.

Φ ¼ tan−1h

r

� �ð2Þ

RESULTS AND DISCUSSION

Quantitative solubility data for carbamazepine in water, poly-ethylene glycol 400, and Na-CMC, PGS, SSG, and HPMCsolid dispersion systems is given in Table 2. Solubility ofcarbamazepine in Na-CMC solid dispersion increased to0.1903 mg/mL from its 0.0061 mg/mL aqueous solubility.

Stepwise multivariate linear regression was performed toevaluate the relationship obtained between response and in-dependent variables. For each polymer 32 full factorial de-sign was applied. The independent and dependent factors,their levels, and the matrix of variables and responses for32 factorial design of each polymer are shown in Table 1.

The statistical evaluation of dependent variables was per-formed by analysis of variance (ANOVA) using MicrosoftExcel Version-2003. The ANOVA results (P value) of thevariables on percentage drug dissolved of solid dispersion

Table 2. Quantitative Solubility Data for Pure Carbamazepine in Different Solid Dispersion Systems and Polyethylene Glycol 400*

Solvents Solid Dispersion System

Solubility of CBZ (mg/mL) Water PEG 400 Na-CMC HPMC PGS SSG

0.0061 0.1782 0.1903 0.1411 0.1168 0.0962

*CBZ indicates carbamazepine; PEG, polyethylene glycol; Na-CMC, sodium carboxymethyl cellulose; HPMC, hydroxypropylmethyl cellulose; PGS,pregelatinized starch; and SSG, sodium starch glycolate.

Sd ¼Pn�1

t¼1 jLog AUCWtð Þ=ðAUCGtð ÞÞjn� 1

ð1Þ

AAPS PharmSciTech 2007; 8 (2) Article 27 (http://www.aapspharmscitech.org).

E4

are shown in Table 3. The detailed summary of results ofregression analysis of SDs for all the polymers is shownin Table 4. The significant parameters in the equations canbe selected using a stepwise forward and backward elim-ination for the calculation of regression analysis. However,in the present study full model having both significant andnonsignificant P values were used in obtaining dependentvariables.23 The coefficients for the equations representingthe quantitative effect of the independent variables on per-centage drug dissolved in distilled water and SGF without

enzymes for each polymer are shown in Table 3. The equa-tions for each polymer can be generated by putting valuesof coefficients in Equation 3.

y ¼ b0 þ b1x1 þ b2x2 þ b11x21 þ b22x

22 þ b12x1x2 ð3Þ

Coefficients with one factor indicate the effect of that par-ticular factor, while the coefficients with more than onefactor and those with second-order terms represent the in-teraction between those factors and the quadratic nature ofthe phenomena, respectively. Positive sign of the term in-dicates positive (additive) effect, while negative sign in-dicates negative (antagonistic) effect of the factor on theresponse.21

It can be concluded from the equations that x1 (amount ofpolymer) showed the largest positive effect, whereas theterm x2 (amount of solvent) showed statistically insignifi-cant positive effect on percentage drug dissolved. The quad-ratic terms of x1 and x2 also had significant positive effecton percentage drug dissolved. The effects of term x1 and x2on particle size of SDs were significant; however, highvalues for standard deviation were observed.

Figure 1 shows the contour plots for percentage drug dis-solved of SDs in distilled water and SGF without enzymesof Na-CMC, SSG, PGS, and HPMC at 30 minutes (Q30),respectively. The contour lines indicated that higher theamount of polymer, the more significant is the dissolutionenhancement. However, for HPMC this was not observed,which may be attributed to controlled release matrix formingability of HPMC.27

The reliability of the equations that described the influenceof factors on percentage drug dissolved was assessed bypreparing 3 additional check points SDs (batch C1, batchC2, and batch C3) in triplicate using the amount of x1 andx2 −0.5, 0.25, and 0.75 level.28 The experimental valuesand predicted values of each response are shown in Table 5.Equation 4 was used to calculate the percentage relative

Table 3. Analysis of Variance Results (P value) Effect of theVariables on Percentage Drug Dissolved of Solid Dispersions*

For Dissolvedin Water

For Dissolvedin SGF

Polymer Factor CoefficientP

Value CoefficientP

Value

Na-CMC X1 12.7 .00 13.2 .00X2 2.12 .01 2.55 .05X1

2 –4.18 .01 –2.15 .23X2

2 –1.30 .15 –1.86 .28X1X2 –0.68 .25 0.15 .89

SSG X1 8.99 .00 9.08 .00X2 1.31 .02 1.62 .02X1

2 –2.21 .02 –2.44 .02X2

2 –0.60 .29 –0.96 .19X1X2 –0.40 .31 –0.30 .51

PGS X1 6.56 .00 6.26 .00X2 1.79 .05 1.26 .03X1

2 –2.39 .10 –2.12 .03X2

2 –0.84 .47 –0.10 .86X1X2 0.63 .44 0.05 .91

HPMC X1 1.76 .15 3.26 .02X2 1.03 .35 1.31 .16X1

2 –4.22 .08 –2.39 .14X2

2 –0.66 .71 –0.02 .99X1X2 –0.30 .81 –0.87 .38

*SGF indicates simulated gastric fluid; Na-CMC, sodiumcarboxymethyl cellulose; SSG, sodium starch glycolate; PGS,pregelatinized starch; and HPMC, hydroxypropylmethyl cellulose.

Table 4. Summary of Results of Regression Analysis of All Polymers*

PolymerResponse: Q30 forDissolution Medium b0 b1 b2 b11 b22 b12

Na-CMC Water 87.13 12.72 2.12 –4.18 –1.30 –0.68SGF 83.25 13.23 2.55 –2.15 –1.86 0.15

SSG Water 78.13 8.99 1.31 –2.21 –0.60 –0.40SGF 76.11 9.08 1.62 –2.44 –0.96 –0.30

PGS Water 76.70 6.56 1.79 –2.39 –0.84 0.63SGF 73.66 6.26 1.26 –2.12 –0.10 0.05

HPMC Water 72.58 1.76 1.03 –4.22 –0.66 –0.30SGF 69.65 3.26 1.31 –2.39 –0.02 –0.87

*Na-CMC, sodium carboxymethyl cellulose; SGF, simulated gastric fluid; SSG, sodium starch glycolate; PGS, pregelatinized starch; and HPMC,hydroxypropylmethyl cellulose.

AAPS PharmSciTech 2007; 8 (2) Article 27 (http://www.aapspharmscitech.org).

E5

error between predicted values and experimental values ofeach response.

% Relative error ¼ jPredicted value � Experimental valuejPredicted value

� �� 100

ð4Þ

The percentage relative error obtained from checkpointbatches was in the range of 0.0637 to 6.6983. Low valuesof the relative error showed that for all the polymers therewas a reasonable agreement of predicted values and experi-mental values. This proved the validity of model and ascer-tained the effects of Na-CMC, SSG, PGS, HPMC and theamount of solvent on percentage drug dissolved.

Particle sizes of pure drug and SDs were determined usingMalvern Mastersizer (Malvern Mastersizer, Worcestershire,UK) with petroleum ether as dispersion medium for sam-ple. Results of particle size analyses are shown in Figure 2.Particle size of pure drug was found to be in a broad range

of 150 to 600 μm. SDs with Na-CMC were observed withmost uniformity in particle size, which ranged from 15 to17 μm. SDs of CBZ with HPMC also showed considerabledecrease in the particle size of CBZ; however, it was spreadover a wide range. In the case of SDs with SSG there were2 particle size distributions first in the range of 138 to158 μm and the other in 1000 to 2500 μm. The latter can bevery well attributed to particle size of plain SSG, whereasfirst can be credited to decreased particle size of drug. In caseof SDs with PGS a small volume fraction was found in therange of 3 to 10 μm and a large volume fraction in the rangeof 100 to 200 μm, whereas in case of SDs with HPMC therewere also 2 distributions but both of them represented drugbecause the difference in particle size was much less. Thisfinding shows that SDs of CBZ with Na-CMC, SSG, PGS,and HPMC showed considerable decrease in the particle sizeof CBZ.

The release of drug from SDs was analyzed in distilledwater and SGF without enzymes. Similarity factor was cal-culated to compare both dissolution profiles for each SD

Figure 1. Contour plots showing (1) percentage drug dissolved in distilled water (solid line), and (2) percentage drug dissolved insimulated gastric fluid (SGF) (dashed line) for solid dispersions prepared with sodium carboxymethyl cellulose (Na-CMC), sodiumstarch glycolate (SSG), pregelatinized starch (PGS), and hydroxypropylmethyl cellulose (HPMC).

AAPS PharmSciTech 2007; 8 (2) Article 27 (http://www.aapspharmscitech.org).

E6

prepared. Table 6 shows the similarity factor values for allthe batches of experimental design of Na-CMC, SSG, PGS,and HPMC polymers. For SDs of all the polymers, the valueof Sd varied between 0.0033 and 0.0139, which is in theproximity of zero. This finding shows that there was no sig-nificant variation between dissolution of SDs in water andSGF without enzymes. In order to assess comparative extent

Table 5. Cross-validation of Model Obtained Using Observed and Predicted Results of Checkpoint Batches*

Polymer Dissolution Media X1 X2 Predicted Values Experimental Values % Relative Error

Na-CMC Water –0.50 0.75 80.85 85.25 5.440.25 0.50 90.70 84.67 6.650.75 –0.50 93.19 93.25 0.06

SGF –0.50 0.75 76.91 80.25 4.340.25 0.50 87.25 88.67 1.620.75 –0.50 90.17 87.65 2.80

SSG Water –0.50 0.75 73.88 75.57 2.290.25 0.50 80.69 82.64 2.410.75 –0.50 82.97 78.15 5.81

SGF –0.50 0.75 71.74 76.36 6.430.25 0.50 78.75 82.37 4.590.75 –0.50 80.60 78.21 2.97

PGS Water –0.50 0.75 73.46 75.68 3.020.25 0.50 78.96 80.26 1.640.75 –0.50 78.93 82.61 4.66

SGF –0.50 0.75 70.87 68.21 3.760.25 0.50 75.71 72.65 4.040.75 –0.50 76.49 78.65 2.82

HPMC Water –0.50 0.75 71.15 68.15 4.220.25 0.50 73.06 70.36 3.700.75 –0.50 70.96 75.28 6.09

SGF –0.50 0.75 68.72 64.12 6.700.25 0.50 70.86 66.29 6.440.75 –0.50 70.42 72.67 3.20

*Na-CMC indicates sodium carboxymethyl cellulose; SGF, simulated gastric fluid; SSG, sodium starch glycolate; PGS, pregelatinized starch; andHPMC, hydroxypropylmethyl cellulose.

Figure 2. Particle size analyses: (A) Pure CBZ, (B) SD ofCBZ with Na-CMC, (C) SD of CBZ with SSG, (D) SD ofCBZ with PGS, and (E) SD of CBZ with HPMC. CBZ indicatescarbamazepine; SD, solid dispersion; Na-CMC, sodiumcarboxymethyl cellulose; SSG, sodium starch glycolate; PGS,pregelatinized starch; and HPMC, hydroxypropylmethyl cellulose.

Table 6. Similarity Factor Calculated for All the SolidDispersions Prepared*

Batches

Similarity Factor (Sd) for Systems

SDs WithNa-CMC

SDs WithSSG

SDs WithPGS

SDs WithHPMC

–1 –1 0.00 0.01 0.01 0.01–1 0 0.01 0.01 0.01 0.01–1 1 0.01 0.01 0.01 0.000 –1 0.01 0.01 0.01 0.010 0 0.01 0.01 0.01 0.010 1 0.00 0.01 0.01 0.011 –1 0.00 0.00 0.01 0.011 0 0.01 0.01 0.01 0.011 1 0.00 0.00 0.01 0.01

*Sd indicates similarity factor; SD, solid dispersion; Na-CMC, sodiumcarboxymethyl cellulose; SSG, sodium starch glycolate; PGS,pregelatinized starch; and HPMC, hydroxypropylmethyl cellulose.

AAPS PharmSciTech 2007; 8 (2) Article 27 (http://www.aapspharmscitech.org).

E7

of dissolution rate enhancement from its SDs, meandissolution time (MDT) was calculated. The dissolutiondata obtained of pure CBZ and SDs of all polymers wastreated according to Equation 5.29

where i is dissolution sample number, n is number of dis-solution sample times, tmid is time at the midpoint betweentimes ti and ti-1, and ΔM is the amount of CBZ dissolved(μg) between times ti and ti-1. The results for MDT calcu-lated are shown in Table 7. The MDT of pure CBZ in waterand SGF without enzymes is 13.55 and 14.07, respectively,which was decreased to 6.83 and 7.94 in the case of SD ofdrug with Na-CMC. This result depicts the fulfillment ofthe objective of dissolution enhancement of CBZ.

The amount of the percentage of drug dissolved in 30 min-utes (Q30) in distilled water and SGF without enzymes forall the SDs prepared according to experimental design isreported in Table 1. SDs of polymers Na-CMC, SSG, andPGS showed an increase in dissolution rate on increase inamount of polymer. Whereas in the case of HPMC soliddispersions, an increase in the amount of polymer up to acertain level, resulted in enhanced dissolution rate but fur-ther addition of polymer resulted in a decrease of dissolutionrate, particularly in water. This finding may be correlated tomatrix-forming ability of HPMC. SDs with Na-CMC didnot show significant difference in dissolution profile carriedin water and SGF without enzymes, whereas SD with SSGshowed significant difference. In case of SD prepared withPGS, increase in amount of polymer showed significantvariation between the dissolution profiles in water and SGFwithout enzymes. Similarity factor (Sd) was determined for

all the experimental batches of SDs with Na-CMC, SSG,PGS, and HPMC. For comparison, the Sd values are shownin Table 6. Figure 3 shows comparative evaluation of dis-solution enhancement efficiency of Na-CMC, SSG, PGS,and HPMC in water and SGF without enzymes.

The SEM images for pure drug and SDs of all polymers areshown in Figure 4. Pure drug image showed crystalline drugof irregular shapes and sizes ranging from 50 to 600 μm,whereas images of SD of drug with Na-CMC up to originalmagnification ×100 did not show any crystalline material.In the case of SDs with SSG, PGS, and HPMC, although asignificant decrease in the size of drug crystals was observed,agglomeration of crystals was also observed.

Figure 5 shows the thermograms for pure CBZ, SDs ofCBZ with Na-CMC, SSG, PGS, and HPMC. Pure CBZshowed a sharp melting endotherm at 194.43-C. A smallendothermic peak at 170.86 was also observed, which canbe attributed to the presence of a small amount of polymor-phic form of carbamazepine. All thermograms except forSD with PGS showed decrease in the energy change of melt-ing endotherm, which confirms a considerable extent of re-duction in crystallinity of drug. In case of thermogram forSD of CBZ with PGS, the melting endotherm for crystallineCBZ disappeared, which ascertains complete amorphizationof CBZ.

The time required for rising water through capillary actionto wet the methylene blue powder was found to be in therange of 3 to 5 minutes for all the solid dispersions, whichwas significantly less when compared with 10 to 15 min-utes for pure drug. For more accuracy in distinguishing wet-tability, contact angle measurement is advised.

For all the SDs, prepared angle of repose was determined.It was found to be in the range of 29- to 35-. This illus-trates the free flowability of SDs and their ability to be usedfor formulation into solid dosage forms.

Table 7. Mean Dissolution Time Calculated for All the Batches of Experimental Design for All Polymers*

Batches

Mean Dissolution Time (MDT)

CBZ: Na-CMC SSG PGS HPMC

Water SGF Water SGF Water SGF Water SGF

–1 –1 8.29 8.50 8.41 9.29 10.4 11.9 9.46 9.29–1 0 7.79 9.46 7.92 10.2 9.62 12.2 9.35 9.22–1 1 8.06 8.79 8.14 9.91 9.83 12.2 10.4 9.190 –1 8.49 9.70 7.98 10.7 9.32 12.3 8.71 8.750 0 8.03 9.03 7.07 8.76 7.80 10.5 8.53 8.690 1 7.78 8.20 6.54 7.85 7.75 9.51 8.36 8.321 –1 7.90 8.44 7.37 7.71 6.57 8.99 7.30 7.551 0 7.40 8.80 6.51 8.23 7.11 9.40 7.07 7.151 1 6.83 7.94 6.34 7.15 6.52 8.58 10.1 9.81

*Na-CMC indicates sodium carboxymethyl cellulose; SSG, sodium starch glycolate; PGS, pregelatinized starch; HPMC, hydroxypropylmethylcellulose; and SGF, simulated gastric fluid.

MDT in vitro ¼Pni¼1

tmid�M

Pni¼1

�Mð5Þ

AAPS PharmSciTech 2007; 8 (2) Article 27 (http://www.aapspharmscitech.org).

E8

Hydrophilic polymer drug solid dispersions increase drugdissolution because of the following possible reasons:

& usually in solid dispersions, the drug is partially dis-solved in melted or dissolved polymer. After dryingof these solid dispersions, the drug will not nucleateto form firm crystals resulting in formation of mi-crocrystals. Drug microcrystals are embedded in thewater-soluble matrix, where hydrophilic polymerspresent the ability of rapid wetting and thereby dis-solution of drug.30 Generally PEGs and PVP soliddispersions follow this principle.

& for solid dispersions of SSG, higher dissolution ratesobserved when compared with other excipients may

Figure 3. Plot showing comparative evaluation of percentage drug dissolved of solid dispersions of sodium carboxymethylcellulose (Na-CMC), sodium starch glycolate (SSG), pregelatinized starch (PGS), and hydroxypropylmethyl cellulose (HPMC)with carbamazepine in distilled water and simulated gastric fluid (SGF).

Figure 4. Scanning electron microscope image of (A) Purecrystalline carbamazepine (CBZ), (B) CBZ: sodium carboxymethylcellulose solid dispersion (SD), (C) CBZ: sodium starchglycolate SD, (D) CBZ: pregelatinized starch SD, (E) CBZ:hydroxypropylmethyl cellulose SD system.

Figure 5. Differential scanning calorimetry thermogramsof (A) Pure crystalline carbamazepine (CBZ), (B) CBZ:hydroxypropylmethyl cellulose solid dispersion (SD), (C) CBZ:sodium starch glycolate SD, (D) CBZ: pregelatinized starch, and(E) CBZ: sodium carboxymethyl cellulose SD system. mWindicates molecular weight.

AAPS PharmSciTech 2007; 8 (2) Article 27 (http://www.aapspharmscitech.org).

E9

be owing to their easy and rapid dispersibility in theaqueous dissolution fluids.31

& solid dispersions of hydrophilic swellable polymerssuch as CMC and HPMC become gelatinized in thedissolution medium. This gelatinized solid dispersionis constantly crushed by the attrition during stirring,and these finely gelatinized SDs diffuse to bulk solu-tion through the diffusion layer.32 Being water reten-tive, gelatinized dispersions also increase wetting ofthe drug, which attributes to increase in dissolution.However, the gelatinized dispersion formed shouldnot be a barrier for the drug diffusion owing to itsviscosity. In the present work, HPMC showed lessdrug dissolution compared with Na-CMC, whichmaybe owing to the formation of highly viscous barrierlayer at the interface of drug and dissolution medium.

CONCLUSION

MLR analysis of results of experimental design illustratedthat SD of carbamazepine when prepared with hydrophilicswellable polymers showed marked increase in percentagedrug dissolution in water and SGF without enzymes, whichwas illustrated with the help of mean dissolution time andsimilarity factor. SDs prepared with Na-CMC showed thehighest drug dissolution compared with HPMC, PGS, andSSG owing to its optimum wetting properties by gelatini-zation and control over particle size of drug. Na-CMC is apopular excipient of tablets and its SD showed good freeflowing properties, which may not need extra excipients forcompression or filling in hard gelatin capsules.

ACKNOWLEDGMENTS

Authors thank All India Council for Technical Education(AICTE) for National Doctoral Fellowship (NDF) and theMetallurgy Department, The M. S. University of Baroda forproviding scanning electron microscopy facility.

REFERENCES

1. Lobenberg R, Amidon GL. Modern bioavailability, bioequivalenceand biopharmaceutics classification system: new scientific approaches tointernational regulatory standards. Eur J Pharm Biopharm. 2000;50:3Y12.

2. Desai KGH, Kulakarni AR, Aminbhavi TM. Solubility of Rofecoxibin presence of methanol, ethanol and sodium lauryl sulphate at(298.15, 303.15 and 308.15) K. J Chem Eng Data. 2003;48:942Y945.

3. Rawat S, Jain SK. Rofecoxib—cyclodextrin inclusion complex forsolubility enhancement. Pharmazie. 2003;58:639Y641.

4. Moneghini M, Voinovich D, Perissutti B, Princivalle F. Action ofcarriers on carbamazepine dissolution. Pharm Dev Technol. 2002;7:289Y296.

5. Bertilsson L. Clinical pharmacokinetics of carbamazepine. ClinPharmacokinet. 1978;3:128Y143.

6. Nokhodchi A, Javadzadeh Y, Siahi-Shadbad MR, Barzegar-Jalali M.The effect of type and concentration of vehicles on the dissolution rateof a poorly soluble drug (indomethacin) from liquisolid compacts.J Pharm Pharm Sci. 2005;8:18Y25.

7. Spireas S, Jarowski CI, Rohera DI. Powdered solution technology:principles and mechanism. Pharm Res. 1992;9:1351Y1368.

8. Javadzadeh Y, Siahi-Shadbad MR, Barzegar-Jalali M, Nokhodchi A.Enhancement of dissolution rate of piroxicam using liquisolid compacts.Farmaco. 2005;60:361Y365.

9. ABBOTT GmbH and Co. Soliqs: nanomorph technology. AccessedNovember 5, 2005.

10. Rasenack N, Hartenhauer H, Müller B. Microcrystals for dissolutionrate enhancement of poorly water-soluble drugs. Int J Pharm. 2003;254:137Y145.

11. Rasenack N, Müller BPR. Dissolution rate enhancement by in situmicronization of poorly water-soluble drugs. Pharm Res. 2002;19:1894Y1900.

12. Sertsou G, Butler J, Scott A, Hempenstall J, Rades T. Factorsaffecting incorporation of drug into solid solution with HPMCP duringsolvent change coprecipitation. Int J Pharm. 2002;245:99Y108.

13. Bakatselou V, Oppenheim RC, Dressman JB. Solubilization andwetting effects of bile salts on the dissolution of steroids. Pharm Res.1991;8:1461Y1469.

14. Valizadeh H, Nokhodchi A, Qarakhani N, et al. Physicochemicalcharacterization of solid dispersions of indomethacin PEG 6000,Myrj 52, lactose, sorbitol, dextrin, and Eudragit E100. Drug DevInd Pharm. 2004;30:303Y317.

15. Chiou WL, Reigelman S. Pharmaceutical applications of soliddispersion systems. J Pharm Sci. 1971;60:1281Y1302.

16. Leuner C, Dressman J. Improving drug solubility for oral deliveryusing solid dispersions. Eur J Pharm Biopharm. 2000;50:47Y60.

17. Ford JL. The current status of solid dispersions. Pharm Acta Helv.1986;61:69Y88.

18. Goldberg AH, Gibaldi M, Kanig JL. Increasing dissolution ratesand gastrointestinal absorption of drugs via solid solutions andeutectic mixtures. II. Experimental evaluation of a eutectic mixture:urea-acetaminophen system. J Pharm Sci. 1966;55:581Y583.

19. El-Zein H, Riad L, El-Bary AA. Enhancement of carbamazepinedissolution: in vitro and in vivo evaluation. Int J Pharm. 1998;168:209Y220.

20. Narang AS, Srivastava AK. Evaluation of solid dispersions ofclofazimine. Drug Dev Ind Pharm. 2002;28:1001Y1013.

21. Hancock BC, Parks M. What is true solubility advantage foramorphous pharmaceuticals? Pharm Res. 2000;17:397Y404.

22. Gohel MC, Panchal MK. Comparison of in vitro dissolution profilesusing a novel, model-independent approach. Pharm Technol. 2000;24:92Y102.

23. Gohel MC, Panchal MK. Novel use of similarity factor f2 and Sdfor the development of diltiazem HCl modified-release tablets usinga 32 factorial design. Drug Dev Ind Pharm. 2002;28:77Y87.

24. Derringer G, Suich R. Simultaneous optimization of severalresponses variables. J Qual Tech. 1980;12:214Y219.

25. Gohel MC, Patel LD. Processing of nimesulide - PEG 400 - PG -PVP solid dispersions: preparation, characterization and in vitrodissolution. Drug Dev Ind Pharm. 2005;29:299Y310.

26. McKenna A, McCafferty DF. Effect of particle size on thecompaction mechanism and tensile strength of tablets. J PharmPharmacol. 1982;34:347Y351.

AAPS PharmSciTech 2007; 8 (2) Article 27 (http://www.aapspharmscitech.org).

E10

27. Cao QR, Choi HG, Kim DC, Lee BJ. Release behavior andphoto-image of nifedipine tablet coated with high viscositygrade hydroxypropylmethylcellulose: effect of coating conditions.Int J Pharm. 2004;274:107Y117.

28. Mashru RC, Sutariya VB, Sankalia MG, Sankalia JM, Parikh PP.Development and evaluation of fast dissolving film of salbutamolsulphate. Drug Dev Ind Pharm. 2005;31:25Y34.

29. Barzegar-Jalali M, Maleki N, Garjani A, et al. Enhancement ofdissolution rate and anti-inflammatory effects of piroxicam using solventdeposition technique. Drug Dev Ind Pharm. 2002;28:681Y686.

30. Arias MJ, Gines JM, Moyano JR, Rabasco AM. Dissolutionproperties and in vivo behaviour of triamterene in solid dispersions withpolyethylene glycols. Pharm Acta Helv. 1996;71:229Y235.

31. Chowdary KP, Rao SS. Investigation of dissolution enhancement ofitraconazole by solid dispersion in superdisintegrants. Drug Dev IndPharm. 2000;26:1207Y1211.

32. Okimoto K, Miyake M, Ibuki R, Yasumura M, Ohnishi N,Nakai T. Dissolution mechanism and rate of solid dispersion particlesof nilvadipine with hydroxypropylmethylcellulose. Int J Pharm.1997;159:85Y93.

AAPS PharmSciTech 2007; 8 (2) Article 27 (http://www.aapspharmscitech.org).

E11